huggingface/transformers

Pull Requests Commits

test on transformers-supported revision

molbap committed 1 year ago

637cadb2

[tests] make cuda-only tests device-agnostic (#35607)

faaany committed 1 year ago

Verified 2fa876d2

[`Compile`] Only test compiling model forward pass (#35658)

ArthurZucker committed 1 year ago

Verified e6f9b034

Enable different torch dtype in sub models (#34873)

zucchini-nlp committed 1 year ago

Verified 84a67891

[`Phi`] bias should be True (#35650)

ArthurZucker committed 1 year ago

Verified 87089176

Removed some duplicated code (#35637)

Sai-Suraj-27 committed 1 year ago

Verified 91f14f1f

Fix whisper compile (#35413)

jiqing-feng committed 1 year ago

Verified b8c34d97

Fix device in rope module when using dynamic updates (#35608)

Cyrilvallez committed 1 year ago

Verified cd44bdb4

Update codeowners with individual model owners (#35595)

Rocketknight1 committed 1 year ago

Verified 15bd3e61

Skip `MobileNetV1ModelTest::test_batching_equivalence` for now (#35614)

ydshieh committed 1 year ago

Verified 1e3c6c1f

Fix flaky `test_beam_search_low_memory` (#35611)

ydshieh committed 1 year ago

Verified 04eae987

Let `EarlyStoppingCallback` not require `load_best_model_at_end` (#35101)

muellerzr committed 1 year ago

Verified b02828e4

Added error when sequence length is bigger than max_position_embeddings (#32156)

Taha1506 committed 1 year ago

Verified 0aaf124f

Use inherit tempdir makers for tests + fix failing DS tests (#35600)

muellerzr committed 1 year ago

Verified 1211e616

Fix flaky `test_custom_4d_attention_mask` (#35606)

ydshieh committed 1 year ago

Verified bbc00046

ArthurZucker committed 1 year ago

f63829c8

[WIP] Emu3: add model (#33770)

zucchini-nlp committed 1 year ago

Verified 52e1f87c

Fix flex_attention in training mode (#35605)

Cyrilvallez committed 1 year ago

Verified ccc0381d

Remove `benchmark.py` after #34275

ArthurZucker committed 1 year ago

a9bd1e62

Chat template: return vectorized output in processors (#34275)

zucchini-nlp committed 1 year ago

Verified e0646f3d

Add Moonshine (#34784)

eustlb committed 1 year ago

Verified 5f087d13

Skip `torchscript` tests if a cache object is in model's outputs (#35596)

ydshieh committed 1 year ago

Verified 6f127d3f

ModernBert: reuse GemmaRotaryEmbedding via modular + Integration tests (#35459)

tomaarsen committed 1 year ago

Verified 6b73ee89

Add flex_attn to diffllama (#35601)

muellerzr committed 1 year ago

Verified 8de7b1ba

ModernBERT bug fixes (#35404)

warner-benjamin committed 1 year ago

Verified 1e3ddcb2

add `_supports_flex_attn = True` for models that do support it (#35598)

ArthurZucker committed 1 year ago

Verified e97d7a5b

[doc] deepspeed universal checkpoint (#35015)

huyiwen committed 1 year ago

Verified c9c682d1

Refactor/fix Cohere2 (#35594)

Cyrilvallez committed 1 year ago

Verified 3a4ae6ea

[`tokenizers`] Ensure that add_prefix_space is propagated to backend_tokenizer.pre_tokenizer (#35593)

tomaarsen committed 1 year ago

Verified 32e0db8a

Fix modular edge case + modular sorting order (#35562)

Cyrilvallez committed 1 year ago

Verified 46276f9a

Older